Keep One's Attention at Recorded Lectures : What effect do embedded questions in recorded lectures have on mind wandering and knowledge gain?

MASTER THESIS

KEEP ONE’S ATTENTION ON RECORDED LECTURES

Researcher Kyra Meutstege

k.meutstege@student.utwente.nl S1962426

Educational Science and Technology

Faculty of Behavioural, Management and Social sciences

First supervisor Dr. Hans van der Meij h.vandermeij@utwente.nl Second supervisor Dr. Henny Leemkuil h.h.leemkuil@utwente.nl

11/09/2019

What effect do embedded questions in recorded lectures have on

mind wandering and knowledge gain?

Acknowledgement

I would like to start with thanking my first supervisor Dr. Hans van der Meij for his patience and guidance over the last year. I have found our meetings just as fascinating and fruitful as his lectures, and I feel like I have learned and progressed a lot since I started this project.

These meetings were even more worthwhile because of Eszter, who read my drafts

numerous times, listened to my ideas, and had immense patience when explaining statistics (again). Thank you for that. I would also like to thank my second supervisor, Dr. Henny Leemkuil for the time and attention to provide very important new feedback and ideas.

Finally, I would like to thank my friends and family for their encouragement and help in good

times and in bad. Without their support and guidance, this research would not have been

possible.

Abstract

Lectures play an increasingly important role in education, for example in MOOCs and Flipped Classrooms. Recorded lectures are however not always as effective as possible, as mind wandering is a common problem for students. Research has shown that mind wandering can have detrimental effects on the learning process of emotional as well as factual nature.

There are some initial indications that embedded questions could reduce the negative effects of mind wandering. Embedded questions could reduce mind wandering by keeping their attention to the literature. This can indirectly lead to knowledge gain, but embedded questions also directly improve knowledge gain through the testing effect. The goal of this study is to research what the effect is of embedded questions in recorded lectures on mind wandering and knowledge gain.

To do so, the current study is a mixed-methods design where quantitative data is supplemented by data from interviews. With a pre- and post-test experiment the effect of embedded questions on mind wandering and knowledge gain are researched. In the current research, no effect was found of embedded questions on either mind wandering or

knowledge gain. Despite the existing consensus, there was also no negative relationship

found between mind wandering and knowledge gain.

Inhoudsopgave

Acknowledgement ... 1

Abstract ... 3

1 Introduction ... 6

2 Theoretical framework ... 8

2.1 Recorded lectures ... 8

2.2 Attention ... 9

2.3 Mind wandering ... 9

2.4 Embedded questions ... 12

2.5 Research questions ... 13

3 Method ... 15

3.1 Research design ... 15

3.2 Respondents ... 15

3.3 Procedure ... 16

3.4 Instrumentation ... 17

3.4.1 Instrumentation for embedded questions ... 18

3.4.2 Instrumentation for measuring of mind wandering ... 18

3.4.3 Instrumentation to measure knowledge gain ... 22

3.5 Data analysis ... 22

3.5.1 Condition ... 23

3.5.2 Mind wandering ... 23

3.5.3 Video engagement ... 24

3.5.4 Knowledge gain ... 24

3.5.5 Relation mind wandering and knowledge gain ... 25

3.5.6 Interview data ... 25

4 Results ... 27

4.1 Mind wandering ... 27

4.2 Video engagement... 29

4.3 Knowledge gain ... 30

4.4 Relation mind wandering and knowledge gain ... 31

4.5 Questionnaire ... 32

4.6 Interviews ... 32

4.6.1 Attention and mind wandering in general ... 33

4.6.2 The experiment ... 34

4.6.3 Attention and mind wandering during the experiment ... 35

4.6.4 Embedded questions and difference between conditions ... 36

5 Discussion and conclusion ... 38

5.1 Mind wandering in a natural situation ... 38

5.2 The effect of embedded questions on mind wandering ... 39

5.3 The effect of embedded questions on knowledge gain ... 40

5.4 The relationship between mind wandering and knowledge gain ... 40

5.5 The effect of embedded questions on mind wandering and knowledge gain ... 41

5.6 Scientific & practical relevance ... 42

5.7 Limitations ... 42

5.8 Future research... 43

6 References ... 45

7 Appendices ... 48

Appendix A: Form reporting mind wandering ... 48

Appendix B: Impression embedded questions ... 49

Appendix C: All embedded questions ... 50

Appendix D: Probe-placement procedure ... 51

Appendix E: Questionnaire ... 53

Appendix F: Interview ... 54

Appendix G: Pre-test ... 55

Appendix H: Post-test ... 56

Appendix I: Factor analysis questionnaire items ... 57

1 Introduction

Lectures are more important in education now than ever (Gorissen, Van Bruggen, &

Jochems, 2012). Although they have played a vital role in the face-to-face classroom for years already, they are now also available as video-recorded lectures. These are not only used to support traditional classes but also as online video lectures which are fundamental elements of MOOCs (Massive Open Online Courses) and Flipped Classrooms (Chen & Wu, 2015). For Flipped Classrooms, the content of the recorded lectures is often used to

generate prior knowledge for the in-class lessons (Herreid & Schiller, 2013). For the MOOCs, the recorded lectures regularly lead to some form of certification (Karsenti, 2013). To ensure that the students gain the knowledge intended, the recorded lectures should be as effective as possible (Schacter & Szpunar, 2015).

One well-known threat to the effectiveness and therefore to the knowledge gain of recorded lectures is the inability of students to keep their attention to the video (Gilboy, Heinerichs, & Pazzaglia, 2015). This inability of the learner to keep the focus on the

important information is also called mind wandering (Corballis, 2012). Research has shown the detrimental effects of mind wandering on the knowledge gain of students. It leads to the fading of information without it having a lasting impact (Risko, Buchanan, Medimorec, &

Kingstone, 2013). This is in line with research by Risko, Anderson, Sarwal, Engelhardt, and Kingstone (2012) who found that students who mind wandered more, performed worse on a retention test afterward. A follow-up study by Risko et al. (2013) found that students who were paying less attention to the lecture (e.g. answering emails whilst listening to the lecture) performed poorer on a subsequent test.

To make recorded lectures more effective, the students should, therefore, be helped to focus their attention on the relevant details of the video. A way to do so could be adding embedded questions. Research has already shown several benefits of embedded questions in video lectures, like a lower in-video dropout (Kovacs, 2016). There are some careful indications that embedded questions can also help with mind wandering (Schacter &

Szpunar, 2015). If embedding questions would be solidly proven to also tackle mind

wandering, it would be a clear and easy suggestion for practitioners to improve their own

recorded lectures.

As shown, to get the best knowledge gain possible from a recorded lecture, students should be helped not to mind wander and to focus their attention. A promising way to do so is the addition of embedded questions. Therefore, the goal of this study is to research if embedded questions can help reduce mind wandering and enhance knowledge gain.

The current study is a mixed-methods design where the quantitative part is adapted from the unpublished study of Szöllősi and Meutstege (2019). They propose an experiment with a pre- and post-test where the effect of embedded questions in recorded lectures on mind wandering and knowledge gain is researched. This quantitative part will be

supplemented with a qualitative part consisting of a questionnaire and interview.

To research whether embedded questions can help reduce mind wandering and enhance knowledge gain, a theoretical framework will be presented after this introduction.

Here, relevant existing literature will be explored. After the theoretical framework, the

method of the current study will be outlined, followed by a detailed description of the

results. After the results section, the discussion and conclusion will follow where the new

results will be compared to the existing literature and possible explanations will be

presented. The last two sections will contain the references and appendices.

2 Theoretical framework

In this chapter, the existing theory relevant for this research will be reviewed. Initially, research about recorded lectures will be explored. The second paragraph will comprise of literature regarding attention which will be closely followed by a paragraph about mind wandering. The chapter will end with a paragraph about embedded questions and their yet researched effect on mind wandering.

2.1 Recorded lectures

Lectures have existed for a long time already and they are now also available as video lectures or recorded lectures (Chen & Wu, 2015). Video lectures have gotten increasingly important through their use in both MOOCs and Flipped Classrooms (Chen & Wu, 2015).

Some experts have said that MOOCs will take higher education to a whole new level, but completion rates are rather low. Not even three per cent of the participants pass the exam at the end (Karsenti, 2013). As recorded lectures are a core feature of MOOCs, it is important to look at the quality of those to guarantee the best learning gain is ensured. Like with the MOOCs, the quality of the video lectures is also very important in the case of Flipped

Classrooms. Students use the recorded lectures to prepare themselves for the work they will do in-class. Educators have already said, however, that it is hard to find videos of good quality (Herreid & Schiller, 2013). Recorded lectures can be enriched with audio and video instruction which is beneficial for the learning experience because people can learn more effectively when both words and pictures are used (Chen & Wu, 2015; Mayer, 2014). But just adding audio and a video does not guarantee sufficient advancement.

Video lessons offer students the autonomy to study at their speed and time. This has many advantages, but also offers challenges. Since the quality of lesson lies in the hands of the students, online video lessons heavily depend on self-regulated learning: “a form of learning in which the learner is primarily responsible for initiating, managing and sustaining the learning process.” (Schacter & Szpunar, 2015, p. 61). According to Randall (2015),

learners who are good at self-regulation should be able to keep their attention more on-task

and should lose their attention less often. Since attention is crucial for effective learning

(Risko et al., 2013), it is important to explore how attention works and how students could

be helped to keep their attention to the learning task.

2.2 Attention

The contemporary world is more and more complex and distraction is everywhere, therefore students need to be able to steer their attention to effectively learn from current learning materials (Risko et al., 2013). During lectures, learners should keep their focus on the important information (D’Mello, 2016; Risko et al., 2013). The amount of research of attention has, however, not even been close to other fields related to learning like

knowledge or actions. Since, as stated above, attention is required for learning, it can be said it is odd that this has not been researched more. What is known, however, is that just

sustaining attention is not sufficient. The limited attentional resources must be effectively distributed by the learner to deal with the dynamic task loads and with the changing learning context. For learning to be effective, the learner has to be able to maintain and properly assign the limited attentional resources (D’Mello, 2016). Attention is needed for cognitive processes like for example the activation of prior knowledge. When attention is lacking, cognitive processes will be hindered (D’Mello, 2016).

D’Mello (2016) states there are four attentional states. A person can either be

attentive or inattentive, which both can be overt or covert. Overt attention would be when a learner is focused on the learning material with both his eyes and his thoughts. When

attention is covert, the learner would think about the learning content, but for an outsider, it looks like he is inattentive. This could, for example, be when the learner has his eyes closed for more concentration. When the learner is overtly inattentive, he is not thinking about the learning content and is off-task. Inattentiveness can, however, also be covert. In that last case, attention drifts away yet it may appear as if the learner is still focused on the task.

Covert inattentiveness is also called mind wandering and research has shown that this can have very detrimental effects (Risko et al., 2012; Risko et al., 2013). For the best learning outcome of a student, a recorded lecture should adequately combat this covert

inattentiveness, or, mind wandering.

2.3 Mind wandering

Learners failing to keep their attention to a task is common, especially during unexciting or

redundant activities. Although the exact amount of mind wandering differs per person and

context, a study by Killingsworth and Gilbert (2010) predicts that people mind wander

around 40% of the time. Everyone has experienced daydreaming about memories or upcoming plans while watching a video and then having to rewind a bit since there is no recollection of what has just been seen or heard (Smilek, Carriere, & Cheyne, 2010;

Stawarczyk, Majerus, Maj, Van der Linden, & D'Argembeau, 2011). Stawarczyk et al. (2011) also call this mind wandering “stimulus-independent and task-unrelated thoughts” (p. 370), or SITUTs. Mind wandering can even occur when the learner is trying hard to keep their attention to the task (D’Mello, 2016).

There is less consensus on the causes of mind wandering. Smallwood (2013) names different hypotheses, namely the executive failure hypothesis, meta-awareness hypothesis, the decoupling hypothesis, and the current concerns hypothesis. The idea that mind

wandering can be caused when control over attention is lost and the learner, therefore, becomes more vulnerable to distraction is called the executive failure hypothesis. The meta- awareness hypothesis suggests that learners are able to recognize when they are mind wandering due to dynamic mental self-monitoring. The decoupling hypothesis suggests that internal and external processes are separated from each other. Therefore, mind wandering (internal) competes autonomously with task performance (external) for attention. The last hypothesis, current concerns (Klinger, Gregoire, & Barta, 1973, in Smallwood, 2013), focuses on the main concern of the mind in combination with the available stimuli and how they influence the thoughts of the learner. This hypothesis suggests that things like yearnings, aspirations and goals which go beyond the perceptual moment can cause mind wandering.

The learners’ thoughts will focus on the most prominent event, which means that if there are not enough stimuli from the lesson, the focus of the learner will shift towards self- generated thought. In other words, the learner will mind wander. A fascinating movie or interesting social interaction, following the current concerns hypothesis, might be

stimulating enough for the learner to keep his/her focus (Smallwood, 2013). This means that

learners will mind wander when 1) the task is not stimulating enough, and/or 2) they have

more prominent off-task stimulation. Since it is outside the scope of the research to study all

hypotheses, the current study will focus on the current concerns’ hypothesis, which is more

relevant now than ever. In the news, there is talk about adults experiencing a lot of stress.

For example, research by a Dutch newspaper Metro

showed that 75% of the youngsters in the Netherlands experience stress because of reasons like ‘insecurity’, ‘pressure on

work/internship’, or ‘too many choices at a too young age’. Another article by a Dutch newspaper, de Volkskrant

, stated that the mental pressure is becoming dangerous and harmful for the health of youngsters. Part of this research will, therefore, explore the current concerns hypothesis.

By enlarge, research shows a negative relationship between mind wandering and learning processes. Although there is some debate about which one is the cause and which the effect, research has shown that people are generally less happy when mind wandering than when they are not (Killingsworth & Gilbert, 2010). Not only does mind wandering have a negative effect of an emotional nature, mind wandering causes negative effects of a factual nature as well. For example, research by Randall, Oswald, and Beier (2014) showed that there is an unfailing negative relationship between mind wandering and on task performance. Not only are there initial findings for mind wandering to be related to unhappiness, but there are more negative sides to mind wandering. For tasks that require nonstop attention, mind wandering can have detrimental effects and cause mistakes (Smilek et al., 2010). Since mind wandering involves an attentional shift from the external

environment to internal thoughts, the learner is no longer attending the important learning content. That in turn occasionally leads to the absence of important knowledge for the learner and result in mistakes (Bixler & D’Mello, 2015; D’Mello, 2016). Research by Risko et al. (2012) showed that learners who mind wandered more often, remembered less about the lecture. When a learner starts to mind wander, it can cause the just learned information to fade. The learning then will not have any long-term effect (Risko et al., 2013).

Bixler and D’Mello (2015) state that the strategies to combat mind wandering can be divided into two categories. The first category is proactive, which refers to strategies

preventing the mind wandering from happening (e.g. mindfulness training). The second category is reactive, which refers to strategies that address the mind wandering while it is

1

https://www.metronieuws.nl/nieuws/dossier/2017/06/longread-waarom-we-ziek-worden-van-drukte Retrieved on 20/08/2019.

2

https://www.volkskrant.nl/nieuws-achtergrond/mentale-druk-op-jongeren-neemt-gevaarlijke-vormen-

aan~bd73895c/ Retrieved on 20/08/2019

happening. This can be done by tailoring the environment so that mind wandering becomes less likely to happen. D’Mello (2016) states that there are some common goals that the approaches to reduce mind wandering share: “(a) capturing attention, (b) giving the learner an opportunity to reflect on the content/activity, and (c) providing an opportunity to correct any comprehension deficits due to mind wandering” (p. 652). Embedded activities who have these goals can combat (the effects of) mind wandering (D’Mello, 2016; Szpunar, Khan, &

Schacter, 2013). Capturing the attention of the learner would show in learners with

embedded questions mind wandering less than learners without embedded questions. The opportunity to reflect for the learner would show in learners who would go back in the video to search for the answer of the embedded question. The opportunity to correct

comprehension deficits would show in a higher learning outcome for participants with the embedded questions as opposed to the learners without embedded questions.

2.4 Embedded questions

Theory shows that embedded questions make the learners retrieve information from

memory. This can cause the ‘testing effect’: long-term memorisation of the learned material.

Empirical research has proven that the testing effect can improve retention of the material more than additional study. For the testing effect to take place, feedback or perfect

performance is not necessary (Roediger III & Karpicke, 2006). So far, instructions often consist of a teacher giving answers or explanations to learners and hence fail to make use of this testing effect. Video lectures can, therefore, improve the learning of the students by adding probes or embedded questions to make them think about the answers/explanations themselves (Williams, 2013). There are also some positive effects found of embedded

questions on mind wandering, but experiments so far have mostly been a first step (Schacter

& Szpunar, 2015; Szpunar et al., 2013). Szpunar et al. (2013) did two experiments to research the effect of embedded questions on mind wandering and knowledge gain. In the first experiment (n = 32), they had a condition where participants received embedded questions after each segment and another condition where students only received a test after

watching all four segments. To measure mind wandering, the research used seven-point

rating scales. In the second experiment (n = 48), Szpunar et al. (2013) replicated their first

experiment with a few changes. The first was the way they measured mind wandering,

which was now through thought-probes: a researcher sat next to the participant and asked

them at random moments whether or not the participant had been mind wandering. The participant then had to write either yes or no on a piece of paper. A second difference was that they added another condition who could restudy the learning content to make sure the learning and attentive benefit came from the embedded questions and not just the re- exposure to the study materials. The study showed that embedded questions in a video lecture can directly lead to less mind wandering which in turn resulted in better learning. It was also shown that the learners that had embedded questions learned more than the students who did not have embedded questions and even learned more than learners who had the chance to restudy the material. Not only did they conclude that embedded

questions can reduce test anxiety, but also that the questions improve learning by helping the learners to combat mind wandering and keep their attention to the video lecture. Since the research of Szpunar et al. (2013) is only, as they say themselves, an initial step, the effect of embedded questions on mind wandering should be explored further, which is the goal of this research.

2.5 Research questions

After exploring the presented theory, some questions are still unanswered. The main aim of this research is to fill in the gaps in existing theory. The main research question is: What is the effect of embedded questions in recorded lectures on mind wandering and knowledge gain?

To answer this research question, the following sub-questions are asked:

1. What is mind wandering like in a real-life situation?

Since this research is the first to study mind wandering in a realistic environment, there are no real expectations.

2. What is the effect of embedded questions in recorded lectures on mind wandering?

Although findings so far have been mostly initial, it is expected based on the study of Szpunar et al. (2013) that embedded questions will reduce mind wandering.

3. What is the effect of embedded questions in recorded lectures on knowledge gain?

Research has shown that embedded questions can elicit the testing effect and can

thus improve retention (Roediger III & Karpicke, 2006). It is therefore expected that

embedded questions will lead to higher knowledge gain.

4. What is the relationship between mind wandering and knowledge gain?

By enlarge, research shows a negative relationship between mind wandering and

learning processes (Bixler & D’Mello, 2015; D’Mello, 2016; Randall et al., 2014; Risko

et al., 2012; Risko et al., 2013; Smilek et al., 2010). It is therefore expected that a

relationship will be found between mind wandering and knowledge gain and that this

will be a negative one.

3 Method

3.1 Research design

The research will be a mixture of both quantitative and qualitative. For the quantitative part, an experiment with an instructional video will be done. The instructional video will be

segmented. The experimental group will receive embedded questions in between the segments and the control group will watch the segmented video without the embedded questions. All participants will receive probes during the video which ask them whether or not they are mind wandering. Before the video, all participants will fill in a questionnaire about mind wandering and after finishing the video, both groups will receive a retention test to study the learning outcomes. For the qualitative part, the students will participate in structured interviews about their mind wandering and how they experienced the video with/without embedded questions. The interviews will be structured since that will help with the comparison between participants (Colton & Covert, 2007).

3.2 Respondents

Both mind wandering and therefore the knowledge gain differs when a video is easy or hard to grasp for the respondent (Smallwood & Schooler, 2006). It is therefore important to match the prior knowledge and learning capability of the participants to the level of content of the video lecture. To ensure that the participants are all more or less equal in this regard, a hard requirement for the level of schooling the participant has had is set. All participants need to have finished at least secondary vocational education. Since the video lesson consists of a lesson ‘Introduction in Law’, it was a prerequisite that the participant had not studied law.

All participation with this experiment will be voluntarily, which means that this sample will not be random. Assignment of the participants to the experimental or control condition, however, will be random as that is the optimal method (Gersten et al., 2005). This means that all participants have an equal chance to be in either the control or the

experimental group. As the goal of this research is to measure group difference (between the control and experimental group), 30 participants per group is deemed sufficient

(VanVoorhis & Morgan, 2007). This means that at least 60 participants were needed in total.

All participants had to give informed consent before participating.

In total there were 61 participants. Three participants had to be excluded from the initial dataset because of problems with the pre- and post-test. Therefore, a dataset of 58 participants was used for calculations, of which 60,3% were female, 37,9% were male and the last 1,7% would not tell. The mean age was 24,60. In terms of prior schooling, 32,8% had been to university level schooling, 36,2% had been to a university of applied sciences, 22,4%

had finished vocational education while 6,9% had finished high school and 1,7% something different. When asked what direction their school was, 22,4% answered healthcare, 20,7%

technical, 15,5% education, 10,3% business, 6,9 economy, 6,9% had a service orientation and 17,2% something else.

3.3 Procedure

The procedure started with potential participants receiving the question of whether or not they wanted to participate via mostly social media (Facebook, Whatsapp, etc.). It was made clear that to participate, the person should have at least finished vocational education and not have studied law.

To ensure a natural environment, participants were asked to sit at a place where they could see themselves studying normally. This could be a desk in a study room or a kitchen table at home. The participants were asked to sit down behind a computer and go to the starting page. A list with supplies was shown which stated that the participant should get the following things: computer or laptop, pen, and the form for the measuring of mind

wandering (see Appendix A: Form reporting mind wandering). Then, the participant had to watch an introduction video where the entire procedure was explained. Important parts here were how to navigate through the website and the explanation of how to report mind wandering. Also, a definition of mind wandering was given, which is the following:

“Having thoughts that have nothing to do with the task you have to carry out. This includes thoughts like: “What shall I eat tonight?”, “I am really busy”, but also: “What is this assignment boring”.

3

Translated from Dutch

Then, the participant had to press a button and they would receive a randomized login number between 1 and 10000. This was done to ensure anonymity and to make it unlikely that people would use the same login name/number. The page ended with a link to the research environment in Graasp, which had a 50/50 chance to go to either the experimental or the control environment to guarantee that every participant had an equal chance to get either condition.

In the research environment, the participant had to log in with the just mentioned randomly generated number. On the first page, they had to sign an informed consent form.

On the next page, the participant filled in the questionnaire about mind wandering, directly followed by the pre-test. After finishing that test, the learner watched the segmented video.

Each segment was on a different page and had to be started manually by the participant. If the participant was part of the experimental group, he/she also had to answer embedded questions in between the segments. After finishing the video, the participant had to fill in the post-test. When the participant was done with this second test, there was a last page with debriefing stating which condition the participant had been part of. The environment ended with several questions regarding whether the participant would participate in the interview and whether or not they were interested in the results of the research, points for the UTwente, and/or VVV-coupons. In total, took every participant about 45 minutes.

If the participant had expressed interest in participating in the interview, the researcher would contact them and schedule a meeting. The interviews took about 10 minutes per interview.

3.4 Instrumentation

In this chapter, the instruments which were used in the research are described. All of the research took place in Graasp. Two different environments were created: one for the control group and another for the experimental groups with the only difference being that the experimental group had embedded questions added after each segment. Graasp was used for this research because of the ease of use and many options it offers.

For the video, a recording of a lesson from the University of Delft was used. The

lesson is Introduction in Law. This video was chosen because it is a good representation of

what a recorded lecture looks like (a combination of lecturer and PowerPoint slides). Also,

since it is introductory, no specific prior knowledge was needed.

3.4.1 Instrumentation for embedded questions

The difference between the control and experimental group is that the latter received embedded questions in between the video segments. Each embedded question was shown directly underneath the video in the format of another video. See Appendix B: Impression embedded questions to see what this looked like on the webpage. This was done so the participant was not able to immediately see the question, but only when watching the embedded question video after finishing the video lecture segment.

Each embedded question was about the content of the segment shown before the question so that it could elicit the before mentioned testing effect. Thus, the question was about already viewed material. As stated before, feedback is not necessary for the testing effect to take place (Roediger III & Karpicke, 2006). To make sure the experimental group did not get any extra information except for the questions themselves, no feedback was added.

Examples of used embedded questions are: “What is jurisprudence?” and “Which jurisdictions are part of public law? And what do these entail?”. For all embedded questions, please see Appendix C: All embedded questions.

3.4.2 Instrumentation for measuring of mind wandering

To measure mind wandering as objectively as possible, the use of physiological measures was considered. Eye-tracking has been used for measuring mind wandering. Most of those researches have however studied eye movements while the participant was reading a text, and not watching a video (Bixler & D’Mello, 2015; Franklin, Broadway, Mrazek, Smallwood,

& Schooler, 2013). The eye movements have proven to be a lot more complex when the participant is watching a video (Mills, Bixler, Wang, & D'Mello, 2016). Research by Mills et al.

(2016) has shown that eye-tracking might potentially be good for studying mind wandering

when watching videos as well. At this moment, however, the relationship between eye

movements and mind wandering while watching a video is not sound enough yet to rely on

for accurate measurement of mind wandering. More research in this field is necessary

because of issues like generalisability of existing models and the accuracy of mind wandering

detection (Bixler & D’Mello, 2014; Mills et al., 2016).

3.4.2.1 Probe-caught method

Since like explained eye-tracking is no option, other instruments will be used in this research.

According to Smallwood and Schooler (2006), experience sampling (or thought sampling) is the most commonly used method to measure mind wandering. Killingsworth and Gilbert (2010) argue that experience sampling is the most reliable method to investigate real-life sentiments. They explain experience sampling as reporting the current thoughts, feelings, and actions of the participant while they engage in an activity. In this research, experience sampling will be used to measure the inner experience of the participant during the experiment (Smallwood & Schooler, 2006). Probe-caught measures, which are a type of experience sampling, will be used. The participants will receive several probes during their viewing of the recorded lecture. Risko et al. (2012) showed that some well-placed probes can reveal interesting attention patterns in connection to mind wandering while disturbing the primal task only minimally. Regarding the probes, an adaptation of the method of Zhao, Lofi, and Hauff (2017) will be used. Like in their research, an auditory signal will be used as a probe. This auditory signal is a certain ‘beep’ that the participants will be familiarized with upfront. During every minute of the video (0:00-1:00, 1:00-2:00 etc.), there is one probe. The specific moment within that time frame will be randomized so that the participants will not be disturbed completely systematically, which is perceived as less interrupting (Bixler &

D’Mello, 2014). However, after randomizing the moment of the probes, they were

sometimes delayed or forwarded if that would disturb the viewing of the video less. For the full procedure of the placement of the probes in the video, please refer to Appendix D:

Probe-placement procedure.

To report the mind wandering, the participant has printed out a form which showed a table to fill in for every segment, like Figure 1. For the full form, see Appendix A: Form reporting mind wandering.

Figure 1. Example of where to report mind

wandering for Segment 1. Translated from

Dutch.

The participant was instructed up front, during the earlier mentioned introduction video, that if he/she had been mind wandering between the former and the current bell, he/she had to write an X under “Yes”. If the participant had NOT been mind wandering, he/she has to put a dash under “No”. The probing process is illustrated in Figure 2.

In the image it is shown that after a probe (auditory signal), the participant has to think: “did I mind wander?” and report with either an X or a –, depending on the answer to the

question. The participant was asked to use the paper for reporting mind wandering during the video because that should 1) be less distracting than having to scroll down during the video to use the online tool and 2) would be more naturally close to for example note- taking. After the participant has watched a whole segment, there was a tool in the

environment where the participant had to copy the data from the on-paper mind wandering report. This tool can be seen in Figure 3.

The participants had to fill in ‘yes’ or ‘no’ for every probe in the video lecture.

Figure 2. An example of mind-wandering reports. Adapted from Zhao, Lofi, and Hauff (2017).

Figure 3. Did you mind wander? Tool for reporting mind wandering in

Graasp.

3.4.2.2 Usage logs

To measure video engagement, a data log was constructed for each video segment. In the columns, the measured variables were shown and in the rows the participants. The following variables were displayed in columns from left to right:

1. Participant number. Every participant was referred to with the randomized login number mentioned in paragraph 3.3 Procedure. These numbers were between 1 and 10000.

2. Condition (coded). Here the condition will be coded with either 0 for the control condition or 1 for the experimental condition.

3. Playtime. This is the total amount of time spent on the video. Plays, replays and pauses are all included.

4. Unique playtime. Here an estimation of how much of the video has been reviewed is shown in percentages. This is done through measuring whether every separate second has been set in “play-mode”.

5. Replay time. In this column, it is shown in percentages how much of the video a participant has played again after seeing the full video. Like the unique play rate, this is measured by checking whether a second of a video has been set in “play-mode”, but this time after the entire video was already finished.

3.4.2.3 Questionnaire and structured interview

Both the questionnaire and the structured interview were used to get insight into the person’s experience with mind wandering. The questionnaire will give more insight into the opinion of the participant, as a questionnaire can measure beliefs and attitudes (Colton &

Covert, 2007). This can help explain possible relationships. The items were structured to a few possible answers with a Likert scale, which makes categorization and with that

comparison of the answers easier (Colton & Covert, 2007). The structured interview is used

to dig deeper into issues that the questionnaire might not uncover (McKenney & Reeves,

2012). The interviews were recorded so that the researcher can first focus on the interview

and later code the answers of the participants. The questionnaire consisted of three types of

questions. The first type asks about effects directly after the video, like “After watching a

video lesson, I soon forget the content”. The second type of questions are about video

lessons themselves. For example: “A video lesson is a good way to gain knowledge for me.”

The third and last type consisted of questions about the feelings/thoughts of the participants during the video, such as “While watching a video lesson, I Mind wander when I am

stressed”. For the full questionnaire, please see Appendix E: Questionnaire.

For the interview, there were several topics the questions focused on. The interviews started with questions about the experiment itself, like “What did you think of your

concentration during the video?”. Other questions were about causes of mind wandering (“When do you mind wander?”), how they combat mind wandering, or what feelings of the participant towards mind wandering are (“If you mind wander, what does that to you?”). For the full interview please see Appendix F: Interview. All interviews were recorded.

3.4.3 Instrumentation to measure knowledge gain

To measure knowledge gain, a pre- and post-test are constructed. The tests were based on the contents of the recorded lecture used for the experiment. Both tests are mostly based on the same topics. Questions are for example “What is meant by the legalistic approach of law? (3 points)” or “Only building plans for dormers with a width of four meters get a permit of the councillor of spatial planning. Is this allowed? Why yes/no? (3 points)”. For the full pre- and post-test, see Appendix G: Pre-test and Appendix H: Post-test. Two questions are repeated and other questions are slightly different. In the pre-test, one question was left out since it required really specific knowledge from the video lectures so it would have been unfair to expect of the participants to know that upfront. This resulted in the pre-test having 10 questions while the post-test consisted of 11 questions.

The tests were analysed with the taxonomy of bloom (Krathwohl, 2002), to make sure not all questions required the same type of cognitive process. Of the total of 11 (post- test) questions, six were ‘Remember’ questions, 2 were ‘Understand’ questions, 2 were

‘Apply’ questions and 1 was an ‘Evaluate’ question. Most questions are thus from the lower parts of the taxonomy. This is deemed appropriate since there is enough variation and the test is about a video of an introductory level.

3.5 Data analysis

In this paragraph, the analysis of the data is described. The paragraph starts with the

quantitative data divided based on the variables, starting with the independent variable

(condition), followed a presentation of the dependent variables (mind wandering, video engagement and knowledge gain). The data analysis ends with the qualitative data, where will be explained how the interviews will be analysed.

3.5.1 Condition

Since the effect of embedded questions cannot be measured directly, it is done indirectly through an experiment with a control group. They both follow the same procedure, with the only difference being the experimental group receiving embedded questions while the control group does not. It was shown by the data that all participants from the experimental group did see the embedded questions. Any differences between the groups are therefore from the embedded questions.

A check on the randomization of participants showed an equal distribution of gender over conditions. However, age was not equally distributed. A one-way ANOVA showed that age was significantly different over conditions F (1, 56) = 4.60, p = .036. Analyses with ANOVAs have been done with and without age as a covariate, but differences in the outcomes were small. It is therefore furtherly ignored.

3.5.2 Mind wandering

Like mentioned earlier, for measuring mind wandering the probe-caught method was used.

As stated in 3.4.2.1 Probe-caught method, it gives yes/no data per participant for each probe. These are converted to a dichotomous variable were 0 means ‘yes’ and 1 means ‘no’.

However, some participants did not fill in the same of yes/no’s as there were probes in a segment. For example, for segment 2 a participant might have reported 5 yes/no’s.

However, in segment 2 there were only 4 probes, which means there is one report too many. In this case, it was assumed that the participant reported also for the moment between the last probe and the end of the video. The last report was thus deleted from the data set.

There were, however, also participants who reported too few yes/no’s. For example, for segment 1 a participant might have filled in two 1’s and two 0’s. This means that the participant has reported a total of 4 times for segment one, while in segment 1 there was a total of 5 probes. The mean of the reports in that segment was calculated, in this case, 0.5.

This mean was then used for calculations instead of the missing value.

For each participant, a total was calculated per segment (a total of 5). This total is converted to a percentage since the number of probes differed per segment. Therefore, these percentages are used in repeated-measures ANOVA to see if there is a trend

(difference over time) in mind wandering. Also, a total of all mind wandering per participant is calculated. Due to non-normal distribution, non-parametric Mann-Whitney tests were used to test whether the mind wandering of the participants differed significantly between the control and experimental group. For the analysis of the difference in the total of mind wandering in the mean totals of all mind wandering between conditions, multiple (stepwise) regression is used.

To analyse what mind wandering looks like in a natural situation, the probe-caught measures will be used to see what the mean amount of mind wandering is. The data from the questionnaire will be used to get a general idea of what people’s perception is of how video lectures influence their attention and knowledge gain. First, it is important to know whether or not the questionnaire data is reliable. For reliability testing, question 1, 4, 6 and 7 were reversed. The questionnaire proved to have good reliability of Cronbach’s α = .88.

With the data, a principal axis factor analysis with oblique rotation will be performed to see which factors can be extracted from the questionnaire and how much of the variance they explain. The newly constructed factors will then be used in a multiple linear regression analysis to test whether or not they are related to mind wandering data from the experiment.

3.5.3 Video engagement

To analyse the effect of embedded questions on video engagement, the playtime, replay time and unique playtime will be compared between the control and experimental group.

This data comes from the earlier mentioned video engagement tool in Graasp. These will be compared through multiple (stepwise) regressions. In case of replay time, it is important to also look at frequencies (how many participants used replay and how many didn’t). This will be done through a Mann-Whitney test.

3.5.4 Knowledge gain

To measure the knowledge gain, the pre- and post-test (see Appendix G: Pre-test and

Appendix H: Post-test) will be compared. Before the analysis, the data had to be prepared.

First, a total of points for both pre- and post-test were calculated per participant. Secondly, these totals of points were converted into percentages since the total amount of points to get per test were different. As a last step of preparation, the knowledge gain was calculated by subtracting the percentage of the pre-test from the percentage of the post-test. The pre- test consisted of 10 items, however one was excluded in reliability tests due to it having zero variance. The remaining 9 items had a Cronbach’s alpha of .72. The post-test had a

Cronbach’s alpha of .76. This means that both the pre- and post-test had acceptable

reliability. Half of the items were scored by another research to measure reliability. Cohen’s Kappa showed for the pre-test .747 and for the post-test .589, which is acceptable inter- rater reliability and means that the scoring of the tests has been reliable. A check on the randomization of participants showed an equal distribution of pre-test score over conditions.

SPSS will be used to compare the mean knowledge gained of the control and the experimental group with a repeated-measures ANOVA. When the results from the experimental and control group are compared, conclusions can be made whether the student who watched the video with embedded questions gained knowledge more, less or equal to the student who watched the video without the embedded questions. It will also be

3.5.5 Relation mind wandering and knowledge gain

To see whether mind wandering could influence knowledge gain, a linear regression will be done. For mind wandering, the sum of mind wandering per segment (so the total of mind wandering in the entire video) will be used. For knowledge gain, the earlier mentioned constructed variable will be used.

3.5.6 Interview data

The recordings of the interviews were transcribed for better analysis. These transcriptions

then were studied to see whether there were any similarities and/or differences between

the statements of the interviewees. Since the interviews were semi-structured, the answers

to the questions could be compared directly. However, the transcriptions were also coded

and sorted to see if there were any relevant statements which were maybe not directly

answers to interview questions.

The data from the interviews will be used mostly to analyse what mind wandering is

like in a realistic environment and what the effect is of the embedded questions on mind

wandering.

4 Results

In this chapter, the results of the analyses are shown. First, the quantitative data will be presented starting with data regarding mind wandering which will be followed by the data concerning knowledge gain. The second part of this chapter consists of the qualitative data from the questionnaires and interviews.

4.1 Mind wandering

In this paragraph, the results concerning mind wandering are shown. See Table 1 for an overview of the amount of mind wandering/attentiveness in percentage of the control and experimental group for each segment and in total. A higher number means less mind wandering.

Table 1. Mean of Control and Experimental Group of Mind Wandering per Segment and Total in Percentages.

N Mean SD

MW seg. 1 Control 28 58.09 25.76 Experimental 28 60.18 26.89 MW seg. 2 Control 30 63.61 25.66 Experimental 28 60.71 23.00 MW seg. 3 Control 30 59.77 27.99 Experimental 28 59.64 24.26 MW seg. 4 Control 30 46.67 30.32 Experimental 28 54.46 30.47 MW seg. 5 Control 30 50.75 27.39 Experimental 28 46.90 30.54 Total Control 27 57.00 20.61 Experimental 28 56.38 21.02

The difference in mind wandering between the control and experimental group was analysed per segment. To measure whether there is a difference in mind wandering

between the control and experimental group in the first segment, a Mann-Whitney test was

conducted. The test showed that the amount of mind wandering did not differ significantly

between the experimental (N = 28, M = 29.04) and the control (N =28, M = 27.96) group, U =

407.00, z = .251, p = .802.

To see whether there is a difference in the means of mind wandering between the control and experimental group in the second segment, a Mann-Whitney test was conducted. The amount of mind wandering did not differ significantly between the control (N = 30, M = 30.50) and the experimental (N = 28, M = 28.43) group, U = 390.00, z = -.487, p = .627.

Another Mann-Whitney test was conducted to analyse whether there is a difference in mind wandering between the control and experimental group in the third segment. The test showed that the amount of mind wandering did not differ significantly between the control (N = 30, M = 30.15) and the experimental (N = 24, M = 28.80) group, U = 400.50, z = -.312, p

= .755.

To examine the difference in the fourth segment between the mind wandering of the control and experimental group, again a Mann-Whitney test was done. It showed that the amount of mind wandering did not differ significantly between the control (N = 30, M = 27.70) and the experimental (N = 28, M = 31.43) group, U = 474.00, z = .855, p = .392.

Again, a Mann-Whitney test was conducted, this time to analyse whether there is a difference in mind wandering between the control and experimental group in the fifth segment. The test showed that the amount of mind wandering did not differ significantly between the control (N = 29, M = 29.95) and the experimental (N = 28, M = 28.02) group, U = 378.50, z = -.444, p = .657.

A multiple regression (stepwise) analysis was performed to see whether condition, age or the pre-test score could predict mind wandering. Based on the results of the study (N = 55), the condition is not significantly related to the mind wandering of the participants, r = -.017, p = .451. Age is also not significantly related to mind wandering, r = .140, p = .154, as well as pre-test score, r = .004, p = .488.

To measure whether the amount of mind wandering differs significantly between the

segments, a repeated-measures ANOVA was done. Age was used as a covariate. Mauchly’s

test indicated that the assumption of sphericity has not been violated, C

(9) = 5.106, p =

.825. The results show that the amount of mind wandering did differ significantly between

the segments, F (1, 52) = 2.45, p = .047. This means that over time, participants started to mind wander more.

4.2 Video engagement

In this chapter, the video engagement of the control and experimental group will be compared to analyze whether there are differences or not. For an overview of the video engagement, please see Table 2. The data of the playtime, unique playtime and replay time of all participants are used for this analysis. For definitions of these variables, please see paragraph 3.4.2.2 Usage logs).

Table 2. Video Engagement. A higher number means a higher (re)view time.

N Mean SD

Playtime Control 30 295.73 56.01 Experimental 28 317.33 43.73 Unique

Playtime Control 30 287.83 53.65 Experimental 28 301.61 17.80 Replay Time Control 30 1.80 7.11

Experimental 28 8.40 23.33

A multiple regression (stepwise) analysis was performed to see whether condition, age, the pre-test score, or mind wandering could predict the playtime. Based on the results of the study (N = 55), condition is not significantly related to playtime, r = .219, p = .054. The trend shows that participants from the experimental group spent more time on the videos than participants from the control group, but since it is not significant the condition did not influence the playtime. The relation between age and playtime is significant, with r = -.248, p

= .034. This means that older participants spent less time on the video than younger participants. The other variables are not significantly related to playtime, with the relation between pre-test score and playtime r = -.120, p = .191, and between mind wandering and playtime r = -.035, p = .399.

To see whether condition, age, the pre-test score, or mind wandering could predict the

replay time, another multiple regression (stepwise) analysis was performed. Based on the

results of the study (N = 55), condition is not significantly related to replay time, r = .184, p =

.090. It is shown by the trend that participants from the control group replayed parts of the

video less than participants of the control group, although the condition cannot predict the replay time. The other variables were also not significantly related, with respectively the relation between age and replay time r = -.119, p = .194, between pre-test score and replay time r = -.103, p = .228, and between mind wandering and replay time r = .047, p = .368.

Again, a multiple regression (stepwise) analysis was done to see whether condition, age, the pre-test score, or mind wandering could predict the unique playtime. Based on the results of the study (N = 55), condition is not significantly related to unique playtime, r = .191, p = .081.

The trend, however, shows that participants from the experimental group have reviewed more of the video than participants of the control group. The relationship was not significant between age and unique playtime r = -.208, p = .064, however, the trend showed that older participants reviewed less of the video. The other variables were also not significant related, with respectively the relation between pre-test score and unique playtime r = -.044, p = .374, and between mind wandering and unique playtime r = -.037, p = .395.

In case of replay time also the frequencies are analysed. When comparing conditions, 93.3%

of the participants from the experimental group (N = 28) have replayed parts of the video at some time, while 75% of the participants form the control group (N = 30) used replay.

Overall, 84.5% of the participants did not use replay anywhere in the video. To analyse whether there is a significant difference in replay frequency between the control and

experimental group, a Mann-Whitney test was conducted. The test showed that the amount of mind wandering did not differ significantly between the control (N = 30, M = 31.07) and the experimental (N = 28, M = 27.82) group, U = 373.00, z = -1.30, p = .195.

4.3 Knowledge gain

For knowledge gain, a repeated-measures ANOVA was done to measure whether the score on the pre-test and the score on the post-test differ significantly. Age was used as a

covariate. The assumption of sphericity has not been violated since there are only two

variables. The results show that the score on the pre- and post-test did differ significantly, F

(1, 55) = 68.10, p = .000. This means that the participants have gained knowledge from the

videos.

A multiple regression (stepwise) analysis was done to see whether condition, age, mind wandering, playtime, replay time, or unique playtime could predict the knowledge gain.

Based on the results of the study (N = 55), condition is not significantly related to knowledge gain, r = .019, p = .445. Age and knowledge gain are significantly related, r = -.279, p = .019, which means that older participants gained less knowledge than younger participants.

Unique playtime and knowledge gain are also significantly related, with r = .230, p = .046.

This means that participants who watched more of the video, gained more knowledge. The other variables are not significantly related, with the relation between mind wandering and knowledge gain r = .166, p = .113, between playtime and knowledge gain r = .166, p = .113, between replay time and knowledge gain r = .044, p = .376.

The mean scores of the participants of both the control and experimental groups for the pre- test, the post-test as well as the difference between them can be seen in Table 3. Also included are the scores of both groups combined.

Table 3. Mean Percentage of Test Scores and Difference of Control, Experimental and Total Group.

N Pre-test Post-test Knowledge Gain

M SD M SD M SD

Control 30 12.58 12.72 39.73 16.15 27.15 11.62

Experimental 28 16.52 14.37 44.00 16.59 27.48 15.24

Total 58 14.48 13.57 41.79 16.36 27.31 13.37

4.4 Relation mind wandering and knowledge gain

To analyze whether mind wandering influences knowledge gain, a linear regression analysis

was performed. The regression analysis revealed that the overall model was not significant,

R

= .07, F (1, 32) = 2.29, p = .140. This means that mind wandering does not influence

knowledge gain.

4.5 Questionnaire

A principal axis factor analysis with oblique rotation was conducted on the 12 questionnaire items. For the full table with all data, please see Appendix I: Factor analysis questionnaire items. The variable Immediate Effect was constructed with five items which explained 45.99% of the variance. Current Concerns was constructed with four items which explained 14.67% of the variance. The last variable, General Effect, was constructed with three items which explained 9.96% of the variance. To identify the reliability of the data, Cronbach’s a was calculated. The results of this showed .88 for Immediate Effect, .77 for Current

Concerns, and .80 for General Effect. This means that the reliability of the construct Current Concerns is acceptable while the reliability of Immediate Effect and General Effect are good.

With these newly constructed variables, a multiple linear regression analysis was performed to test whether they can predict mind wandering. The regression analysis, as shown in Table 4, revealed that the overall model was not significant, R

= .11, F (3, 51) = 2.17, p = .103. This means that the variables Immediate Effect, Current Concerns and General Effect cannot be used to predict mind wandering.

Table 4. Linear model of predictors of mind wandering.

b SE b p

Constant 14.10 .69 .000

Immediate Effect -1.27 .87 -.23 .150

Current Concerns -1.05 .80 -.18 .198

General Effect -.04 .95 -.01 .967

Note. R

= .22.

4.6 Interviews

In this chapter, the results of the interviews are shown. Pseudonyms are used to protect the

privacy of the interviewees. Four participants from the control group (Adam, James, Emma

and Sophie) and four participants from the experimental group (George, Emily, Jennifer and

Michael) were interviewed. Both groups included two males and two females. The youngest

interviewee was 23 and the eldest 54. The interview results are shown starting with a more

general view on attention and mind wandering in general (4.6.1) Then the results will dig deeper into these topics concerning the experiment itself (4.6.2) and the attention and mind wandering during the experiment (4.6.3). The chapter ends with a paragraph about the opinion of the participants about the embedded questions (4.6.4).

4.6.1 Attention and mind wandering in general

Almost all participants mentioned that it is very important for their concentration to have a genuine interest in the content. One participant added that for him the goal in a lecture is to gain knowledge. “If I understand something and then a very boring example is given or they have to explain it again … then I mind wander.” Mind wandering can be caused through something external, like Adam stating that social media can be a big distractor for him.

Causes can also be more internal and based on people’s current concerns. While one mentioned that she mind wanders more often when she feels stressed, another added that she mind wanders when she is tired or when she is busy and thinks about all the things she needs to do. Michael stated that he could imagine that people would mind wander more when they have a lot of things going on, however, he does not recognize that in own experiences. George’s answers are in line with Michael’s: “I would say, logic would lead to me saying yes [more mind wandering when having a lot on your mind], however, I cannot connect this to my experience from the video”. This could be explained by Emma, who stated that current concerns like “stress” or “sleeplessness” might influence the amount of mind wandering, but doesn’t think it is as big of an influence as the earlier mentioned interest in the content.

To battle mind wandering, it differs what the participants do. Some are aware of their attention already during their scheduling. Both Emma and Michael state that they have better concentration in the morning and therefore often plan their study time then. Emma admits that even then she could lose her concentration, but those were her more

“productive” hours. Others have measures for during the studying itself. George states that

to have good concentration, he needs to stay hydrated. Adam told about some precautions

he takes against mind wandering: “You see my desk here, I will describe it. I put my desk

against the wall with hardly any external stimuli around me. …. I have to really isolate

myself.” Since he was aware that social media is one of his biggest distractions, he also had

an app installed on his phone with which he could disable everything on his phone for an

hour. To keep her attention on the lecture, Sophie types along with the lesson. She noticed that it was harder to concentrate during this experiment because she didn’t type with the lecture this time. There is also more general prevention of mind wandering by practicing.

Jennifer has been more aware of her mind wandering lately and to battle it, she tries to practice keeping her focus on the task for a longer period. She states that she experiences improvements in for example awareness: “I am more aware of ‘oh, I am not thinking about that anymore’. Then it is easier to say to yourself like: ‘you have to go back’.”

The general feelings of participants about mind wandering were sometimes negative, as one of them mentioned to have felt like: “What have I been doing here [in a lecture], I have only been mind wandering.”. Another participant even looked at mind wandering as

something that results in a punishment, by saying “The punishment is that you have to go home and in case of a lecture you have to read your book a bit more attentive”. Although not in a lecture environment, Jennifer expressed that while she feels neutral about mind

wandering in general, it does bother her when she is talking to other people and misses parts of the conversation.

However, despite these negative sides to mind wandering the overall feeling connected to mind wandering was, for most participants, neutral. This showed in statements such as

“To be honest, I don’t really care” and “What [mind wandering] does to me? Not much”. In some cases, the feelings toward mind wandering were even positive. James answered that he found mind wandering okay. He stated the following: “I think [mind wandering] is okay. I don’t miss any information that I would have wanted to know. Or, in principle, yes, but I would not have found it interesting.”. Liam even quantified mind wandering as being

‘relaxed’. “I don’t really mind it at all.” The feelings mentioned in this research are therefore not in line with the findings from the research mentioned above.

4.6.2 The experiment

Most of the participants stated that they found the video lesson interesting. While the

content was new for some, for others it was more of a refresher or repetition since they

already had learned some from history lessons during middle school of from television

shows. Especially for the participants for whom the information was unfamiliar, the lesson

had some parts that were harder to follow. Jennifer even said that she “kind of dropped out

of the lesson when the content was really hard to understand”.